Tab 1

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

DATA SCIENCE PROJECT

AIM: Applying Statistics(DESCRIPTIVE & INFERENTIAL) on an BANK DATASET

with 2000 rows and 9 columns(Acocunt_ID, Customer_name, Account_type,
Balance, Branch, Status, Created_At, Loan_Amount, Interest_Rate) and applying
various statistical formulas to it.

Here are top 100 data out of Dataset:

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

CSV File:bank_system_data.csv

DESCRIPTIVE STATISTICS:
Here I have applied the descriptive statistics (mean, median, mode, IQR, variance,
standard deviation, and range) to the top 100 data from the csv file I uploaded.

CODE:

import pandas as pd

path="/content/drive/MyDrive/bank_system_data.csv"
df = pd.read_csv(path)

df.head(100)
try:
df = pd.read_csv('bank_system_data.csv')
display(df.head())
print(df.shape)
except FileNotFoundError:
print("Error: 'bank_system_data.csv' not found.")
df = None
except pd.errors.ParserError:
print("Error: Could not parse the CSV file.")
df = None
except Exception as e:
print(f"An unexpected error occurred: {e}")
df = None
import numpy as np

# Calculate descriptive statistics for the top 100 data points

top_100_df = df.head(100)

# Numerical features

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

numerical_features = ['Balance', 'Loan_Amount', 'Interest_Rate']

# Descriptive statistics
for feature in numerical_features:
print(f"Descriptive statistics for {feature}:")
print(f"Mean: {np.mean(top_100_df[feature])}")
print(f"Median: {np.median(top_100_df[feature])}")
print(f"Mode: {top_100_df[feature].mode()[0]}")
print(f"IQR: {np.percentile(top_100_df[feature], 75) - np.percentile(top_100_df[feature],
25)}")
print(f"Variance: {np.var(top_100_df[feature])}")
print(f"Standard Deviation: {np.std(top_100_df[feature])}")
print(f"Range: {np.max(top_100_df[feature]) - np.min(top_100_df[feature])}")
print("-" * 20)

OUTPUT:

Descriptive statistics for Balance:

Mean: 53137.34709999999
Median: 57411.595
Mode: 109.4
IQR: 40903.7225
Variance: 719679162.2630583
Standard Deviation: 26826.836605590648
Range: 99366.68000000001
--------------------
Descriptive statistics for Loan_Amount:
Mean: 29046.043400000002
Median: 32963.365
Mode: 35.68
IQR: 21746.47
Variance: 209940534.04584235
Standard Deviation: 14489.324830572416
Range: 49421.15
--------------------
Descriptive statistics for Interest_Rate:
Mean: 7.2698
Median: 7.655
Mode: 3.02
IQR: 4.420000000000002
Variance: 6.31287996
Standard Deviation: 2.5125445190085687
Range: 8.81
--------------------

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

SKEWNESS AND KURTOSIS:

CODE:
import pandas as pd
import numpy as np
import scipy.stats as stats
import matplotlib.pyplot as plt
import seaborn as sns

# Assuming 'df' is your DataFrame and 'Balance', 'Loan_Amount', 'Interest_Rate' are your
numerical features
numerical_features = ['Balance', 'Loan_Amount', 'Interest_Rate']

def analyze_skewness_kurtosis(data, feature):

"""
Calculates and prints skewness and kurtosis, and displays a histogram.

Args:
data: The pandas DataFrame containing the data.
feature: The name of the numerical feature to analyze.
"""
skewness = stats.skew(data[feature])
kurtosis = stats.kurtosis(data[feature])

print(f"Analysis for {feature}:")

print(f"Skewness: {skewness:.2f}")
print(f"Kurtosis: {kurtosis:.2f}")

# Interpretation of skewness
if skewness > 0:
print("Distribution is right-skewed (positive skew).")
elif skewness < 0:
print("Distribution is left-skewed (negative skew).")
else:
print("Distribution is approximately symmetrical.")

# Interpretation of kurtosis
if kurtosis > 0:
print("Distribution is leptokurtic (heavy-tailed).")
elif kurtosis < 0:
print("Distribution is platykurtic (light-tailed).")
else:
print("Distribution is mesokurtic (normal-tailed).")

# Display histogram
plt.figure(figsize=(8, 6))
sns.histplot(data[feature], kde=True)

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

plt.title(f"Distribution of {feature}")
plt.xlabel(feature)
plt.ylabel("Frequency")
plt.show()
print("-" * 30)

# Apply the analysis to each numerical feature

for feature in numerical_features:
analyze_skewness_kurtosis(df, feature)

OUTPUT:
Analysis for Balance:
Skewness: -0.02
Kurtosis: -1.16
Distribution is left-skewed (negative skew).
Distribution is platykurtic (light-tailed).

------------------------------
Analysis for Loan_Amount:
Skewness: -0.08
Kurtosis: -1.21
Distribution is left-skewed (negative skew).
Distribution is platykurtic (light-tailed).

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

Analysis for Interest_Rate:

Skewness: 0.00
Kurtosis: -1.24
Distribution is right-skewed (positive skew).
Distribution is platykurtic (light-tailed).

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

INFERENTIAL STATISTICS:
Here I have applied the python algorithm to represent LINEAR REGRESSION
graphically.

CODE:

import pandas as pd
from scipy import stats
import statsmodels.api as sm
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.linear_model import LinearRegression
import numpy as np

# Load your dataset (or reuse existing DataFrame)

# df = pd.read_csv('your_file.csv')
df_100 = df.head(100) # First 100 rows

# 1. One-Sample T-Test: Is avg loan amount different from $25,000?

t_stat, p_val = stats.ttest_1samp(df_100['Loan_Amount'], 25000)
print("One-sample t-test (Loan_Amount vs $25,000):")
print(f"t-statistic = {t_stat:.3f}, p-value = {p_val:.3f}\n")

# 2. Two-Sample T-Test: Comparing average balances between two account types

savings = df_100[df_100['Account_Type'] == 'Savings']['Balance']
checking = df_100[df_100['Account_Type'] == 'Checking']['Balance']
t2_stat, p2_val = stats.ttest_ind(savings, checking, equal_var=False)
print("Two-sample t-test (Savings vs Checking - Balance):")
print(f"t-statistic = {t2_stat:.3f}, p-value = {p2_val:.3f}\n")

# 3. ANOVA: Compare loan amounts across account types

anova_groups = [group["Loan_Amount"].values for name, group in
df_100.groupby("Account_Type")]
f_stat, anova_p_val = stats.f_oneway(*anova_groups)
print("ANOVA (Loan_Amount ~ Account_Type):")
print(f"F-statistic = {f_stat:.3f}, p-value = {anova_p_val:.3f}\n")

# 4. Chi-Square Test: Account_Type vs Status

contingency_table = pd.crosstab(df_100['Account_Type'], df_100['Status'])
chi2, chi_p, dof, expected = stats.chi2_contingency(contingency_table)
print("Chi-square test (Account_Type vs Status):")

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

print(f"Chi2 = {chi2:.3f}, p-value = {chi_p:.3f}, dof = {dof}\n")

# 5. Linear Regression: Predict Loan Amount from Balance

X = df_100[['Balance']]
y = df_100['Loan_Amount']
model = LinearRegression().fit(X, y)
print("Linear Regression (Loan_Amount ~ Balance):")
print(f"Intercept: {model.intercept_:.2f}, Slope: {model.coef_[0]:.4f}")
r2 = model.score(X, y)
print(f"R-squared: {r2:.3f}")

# Optional: Plot regression

plt.figure(figsize=(8, 5))
sns.regplot(x='Balance', y='Loan_Amount', data=df_100)
plt.title('Linear Regression: Loan Amount vs Balance')
plt.tight_layout()
plt.show()

OUTPUT:

One-sample t-test (Loan_Amount vs $25,000):

t-statistic = 2.778, p-value = 0.007

Two-sample t-test (Savings vs Checking - Balance):

t-statistic = -1.192, p-value = 0.237

ANOVA (Loan_Amount ~ Account_Type):

F-statistic = 0.325, p-value = 0.723

Chi-square test (Account_Type vs Status):

Chi2 = 1.924, p-value = 0.750, dof = 4

Linear Regression (Loan_Amount ~ Balance):

Intercept: 33819.62, Slope: -0.0898
R-squared: 0.028

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

Machine Learning Algorithm:

—>Logistic Regression:

CODE:

# STEP 1: Import necessary libraries

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import classification_report, confusion_matrix, accuracy_score

# STEP 2: Load your dataset

# If already loaded, skip this. If not, upload and load:
# from google.colab import files
# uploaded = files.upload()
# df = pd.read_csv("your_file.csv")

# STEP 3: Use first 100 rows

df_100 = df.head(100).copy()

# STEP 4: Check column names to avoid typos

print(df_100.columns)

# STEP 5: Encode target variable

df_100['Status'] = df_100['Status'].map({'Active': 1, 'Inactive': 0})

Prepared by:Harsh Bhavsar

IU2241230442 Data Science – CE0630 PROJECT 6-CSE-C2

# STEP 6: Handle missing values (if any)

df_100 = df_100.dropna(subset=['Balance', 'Loan_Amount', 'Interest_Rate', 'Status'])

# STEP 7: Define features and target

X = df_100[['Balance', 'Loan_Amount', 'Interest_Rate']]
y = df_100['Status']

# STEP 8: Train-test split

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

# STEP 9: Train logistic regression model

model = LogisticRegression()
model.fit(X_train, y_train)

# STEP 10: Predict and evaluate

y_pred = model.predict(X_test)

print("\nConfusion Matrix:\n", confusion_matrix(y_test, y_pred))

print("\nClassification Report:\n", classification_report(y_test, y_pred))
print("\nAccuracy Score:", accuracy_score(y_test, y_pred))

OUTPUT:

Faculty Name: Dr. Sheetal Pandya

Signature:___________________

Prepared by:Harsh Bhavsar